Made the first argument to uip_connect() const, to ensure it isn't altered
Importing patch daef1ea252feca2ae540c2c06f5f49496231a209 from Contiki.
Original commit by Adam Dunkels <adam@thingsquare.com>
Change-Id: I92a8c389b0331c2cddcdac0190965209a4488eb3
Signed-off-by: Jukka Rissanen <jukka.rissanen@linux.intel.com>
Added comment how UDP packet is handled when receiving it.
Change-Id: If1e10c5bde55cb2cf7a8928b0f1ec38dd6ab84b4
Signed-off-by: Jukka Rissanen <jukka.rissanen@linux.intel.com>
Make the debug prints a bit more useful by printing also
the relevant IP addresses.
Change-Id: I2029f7a7bc5e888d184467dee19802af9621e36e
Signed-off-by: Jukka Rissanen <jukka.rissanen@linux.intel.com>
Disabling prints by default but have them around as we will need
them later.
Change-Id: I21d980018286be560b9a072ee08ef0be42b0e788
Signed-off-by: Jukka Rissanen <jukka.rissanen@linux.intel.com>
We have an array of UDP connections so just go through
the list in a most simple way. Set only the UDP connection
pointer if we found something and clear it otherwise.
This helps to avoid weird errors where multiple UDP connections
were added but only the first one was found during list traversal.
Change-Id: Iae90ee6803eee1d06e7a49211d77692b0cf5ca50
Signed-off-by: Jukka Rissanen <jukka.rissanen@linux.intel.com>
It is better to initialize the UDP connection array
fully in init instead of setting only the port number
to zero. This way we do not have any garbage bytes
left around.
Change-Id: I90dcf6dc39a5a108818f25796160ca7dc52ba9d9
Signed-off-by: Jukka Rissanen <jukka.rissanen@linux.intel.com>
The uip6.c:uip_process() was not checking the upper bound of
the UDP connection list when traversing the connection list.
Now the code checks the upper bound of the array (UIP_UDP_CONNS)
and will not go beyond the max number of items in the array.
Change-Id: I417ffd2bbb54922ae071f2f382f7e9470516ce55
Signed-off-by: Jukka Rissanen <jukka.rissanen@linux.intel.com>
This fixes RPL compile errors if CONFIG_NETWORKING_WITH_RPL is
enabled. The end result is not tested atm so RPL might not yet
work as expected.
Change-Id: I458c20f8f8c7c750cc17f24acf425f244399699d
Signed-off-by: Jukka Rissanen <jukka.rissanen@linux.intel.com>
The process: label is only used if we are a router so add
ifdef guard.
Change-Id: Ib88d4e4d70b049541480a6e9dd9d68911543ff50
Signed-off-by: Jukka Rissanen <jukka.rissanen@linux.intel.com>
If the udp packet was successfully delivered to the application,
and if the application did not want to send anything back to
sender, then then uip_slen(buf) will be 0.
In this case we MUST NOT set uip_len(buf) to 0 as that would
cause the net_buf to be released by RX fiber. In this case it is
application responsibility to release the buffer.
Change-Id: I43cf9b147a45294288dc3c631908b034e1a4ab6a
Signed-off-by: Jukka Rissanen <jukka.rissanen@linux.intel.com>
Make tcpip_input() to return 0 if there is an error when
processing the packet. Value != 0 indicates a successfull
packet processing and in this case the application will
free the packet.
Change-Id: I4aaeb5f0039cfbe25f7238a144964e31a8204148
Signed-off-by: Jukka Rissanen <jukka.rissanen@linux.intel.com>
When read from the Rx FIFO, pass the network buffer to the uIP IPv6
stack using the tcpip_input() function. As the event is handled, not
all cases where the buffer is to be discarded can be detected.
Thus, if ICMPv6 errors will not be sent, indicate that the packet is
to be dropped by setting its length to zero.
Change-Id: I49f69011685e39759ba25ebb07670308382a5ca8
Signed-off-by: Patrik Flykt <patrik.flykt@linux.intel.com>
Signed-off-by: Jukka Rissanen <jukka.rissanen@linux.intel.com>
Fixed the compiler warning about not returning correct value to
caller. Now we return correct value if neighbor solicitation was
sent properly.
Change-Id: Ie28a2ee417bcdf3eb9b24ab91a49c688992e7420
Signed-off-by: Jukka Rissanen <jukka.rissanen@linux.intel.com>
This is enabled so that the loopback network test
driver works.
Change-Id: Ib5e1af2c1b064b0ef71c31be919ea88ed62531ff
Signed-off-by: Jukka Rissanen <jukka.rissanen@linux.intel.com>
This mega patch changes how Contiki uses buffers. In standard
Contiki there is only couple of static and global buffers.
This makes the stack non-reentrant and only usable in one
thread. This patch replaces the global buffer with buffer
supplied by caller (net_buf.h). The Contiki stack is still not
fully re-entrant after this (packet reassembly needs more TLC)
but it is a good start.
Change-Id: I63abc230b36b14f33f687d3ef64cffb0f3a69f5d
Signed-off-by: Jukka Rissanen <jukka.rissanen@linux.intel.com>
Base is from Contiki git tree at git://github.com/contiki-os/contiki.git
commit 6ac939bc58ea76aa37e9b86af55ad51cfc3434e9
This one does not do anything useful yet. It sets the baseline for
subsequent patches.
Change-Id: I7078c938dd59188beceaf47721afaa521d8bde1a
Signed-off-by: Jukka Rissanen <jukka.rissanen@linux.intel.com>
